Cloud Firestore Pricing 2026

Cloud Firestore's free Spark Plan includes 50,000 document reads per day, 20,000 document writes per day, 20,000 document deletes per day, 1 GiB of stored data, and 10 GiB of outbound data transfer per month. These quotas reset daily at midnight Pacific time and apply only to the default database. Named (non-default) databases, TTL deletes, point-in-time recovery (PITR) data, backups, and restore operations all require billing to be enabled regardless of usage volume.

When you attach a realtime listener, you are charged for a read each time a document in the result set is added or updated, and when a document is removed from the result set because the document changed (not for deletions). Additionally, if offline persistence is enabled and the listener is disconnected for more than 30 minutes, you are charged for all documents as if issuing a brand-new query. If offline persistence is disabled, every reconnect — even brief network interruptions — triggers a full re-read charge.

No — unlike most GCP services, Firestore read and write operations do not support configurable quota limits. You cannot set a hard cap that stops Firestore from processing operations once a threshold is reached. The only protection available is setting GCP billing budget alerts, which send notifications but do not stop usage. The recommended approach is to design your app to avoid runaway patterns (no infinite listeners, no unbounded loops that trigger writes) and monitor Cloud Monitoring metrics for anomalous spikes.

For small apps that stay within or near the free Spark plan, Firestore is typically much cheaper than managed SQL databases. A managed PostgreSQL instance (e.g., Cloud SQL) carries a fixed monthly cost for a production instance, while Firestore costs $0/month on the Spark (No-cost) plan for apps with modest usage and well-optimized queries. The cost advantage narrows at high scale or with access patterns that generate excessive reads.

Several patterns significantly inflate Firestore costs: (1) SQL-style normalized data models where each user action requires multiple separate document reads, (2) Realtime listeners attached to large or frequently-updated collections, (3) Security rules that read additional documents for authorization checks, (4) High-frequency small write workloads such as IoT telemetry, per-event analytics, or chat messages, and (5) Using multi-region database configuration when single-region would suffice at half the price.

At small scale, apps with well-optimized queries can often run on the Spark (No-cost) plan at $0/month. Beyond that, the Blaze (Pay as you go) plan bills by usage and has no published flat price, so cost depends on the number of reads, writes and stored data. Costs scale with usage, and many apps run entirely within the free tier.

It depends on the workload. A Firebase team member confirmed that Firestore has lower per-GiB storage costs, making it cheaper for storage-heavy apps. However, for high-frequency small-write workloads — such as IoT devices writing a few times per second — Firestore's per-operation pricing can be significantly more expensive because it charges per document operation rather than by data size.

The Spark (No-cost) plan includes 50,000 document reads per day, 20,000 document writes per day, 20,000 document deletes per day, 1 GiB of stored data, and 10 GiB of outbound data transfer per month. These quotas apply only to the default database — named databases have no free quota and require billing from the first operation. Features like backups, PITR, and TTL deletes also require a billing account regardless of other usage.

Yes — the Spark (No-cost) plan is permanently free within daily limits, and the Blaze (Pay as you go) plan preserves those same free-tier quotas. However, Cloud Functions (commonly used with Firestore) require a billing account to be attached even to access their own free tier. Additionally, enabling minInstances on Cloud Functions to avoid cold starts negates the scale-to-$0 benefit, as idle instances are billed continuously.

EU compliance is an active concern for some developers. At least one developer reported having lawyers review EU legislation before committing to Firebase as a backend. Google offers EU multi-region database locations for Firestore, but whether this satisfies specific GDPR obligations depends on your data processing agreements and the nature of personal data involved — legal review is advisable.

You'll need to upgrade to the Blaze (Pay as you go) plan when your app exceeds any of the Spark daily limits (50,000 reads, 20,000 writes, 20,000 deletes, or 1 GiB stored data), when you need a named (non-default) database, when you require backups or point-in-time recovery, when you use TTL deletes, or when you need to deploy Cloud Functions (which require an active billing account even if you stay within the Functions free tier).

Yes. Reading Firestore data through the Firebase console — such as browsing collections and documents — counts toward your daily document read quota on the Spark plan. For apps near the 50,000 reads/day limit, avoid routine console browsing during peak usage periods.

The two most commonly cited hidden costs are: (1) per-operation charges that spike for write-heavy workloads like IoT or real-time event streams, and (2) Cloud Functions minInstances costs to mitigate cold starts, which are billed per function and eliminate the scale-to-zero benefit. Additionally, the free Spark tier's daily quotas apply only to the default database—named databases require billing from the first byte used.

The Spark (No-cost) plan includes 50,000 reads/day, 20,000 writes/day, 1 GiB storage, and 10 GiB outbound transfer per month at no charge forever. On the Blaze (Pay as you go) plan, idle apps incur near-zero costs. However, some features require a billing account even when actual spend is $0—including Cloud Functions, named databases, and backups.

Cloud Firestore on the Blaze (Pay as you go) plan charges per document operation (reads, writes, and deletes) as well as for data stored and outbound transfer. It is not primarily a storage-based pricing model. For workloads with many small frequent writes—such as IoT devices checking in multiple times per second—the per-operation cost can be significantly higher than storage-based alternatives like Firebase Realtime Database.

No. The free daily quota applies only to the default database. Named (non-default) databases require a billing account on the Blaze plan and are billed from the first byte of usage, with no free quota.

Yes. Deploying Cloud Functions requires upgrading to the Blaze (Pay as you go) plan and attaching a valid payment method, even if your Functions usage stays within Google's free tier limits for Functions. You will not be charged if you remain within free tier thresholds, but a billing account is required to deploy Functions at all.

It depends on write frequency. Firestore's per-operation pricing can become significantly more expensive than alternatives for IoT workloads where devices send small payloads many times per second—you accumulate high write operation counts while storing very little data. For these patterns, compare Firestore's operation costs against Firebase Realtime Database or other storage-priced alternatives before committing.

You will need to upgrade to the Blaze (Pay as you go) plan if your app exceeds 50,000 reads/day, 20,000 writes/day, or 1 GiB of stored data; if you use named databases; if you need backups, point-in-time recovery, or TTL deletes; or if you deploy Cloud Functions. Browsing Firestore data in the Firebase console also counts toward your daily read quota.

Yes — the Spark (No-cost) plan is permanently free, not a time-limited trial. It includes 50,000 document reads/day, 20,000 document writes/day, 20,000 document deletes/day, 1 GiB stored data, and 10 GiB outbound transfer/month. Note that free quota applies only to the default database; named databases require the Blaze (Pay as you go) plan from the first byte of usage.

No credit card is required for the Spark (No-cost) plan. However, adding Cloud Functions to your Firebase project requires upgrading to the Blaze (Pay as you go) plan and attaching a billing account — even if your total usage remains within the free-tier thresholds for both Firestore and Functions.

Firestore works for IoT, but its operation-based pricing model can make it expensive for fleets writing at high frequency. A fleet of IoT devices checking in a few times per second generates large numbers of write operations despite small payload sizes—potentially making Firestore more expensive than storage-priced alternatives. Evaluate your total expected monthly write operations before selecting Firestore for write-intensive IoT workloads.